The coat protein of a virus represents a component of the virion that provides physiochemical stability and protection to the enclosed viral genome. Enveloped viruses contain a host-derived lipid membrane that surrounds the coat protein (as observed in archaeal viruses of the filamentous Lipothrixviridae), while other viruses contain an internal lipid membrane. STIV is one such inner lipid membrane-containing virus that belongs to an unclassified viral family within Crenarchaeota. The crystal structure of the major coat protein (MCP) from STIV (B345) reveals a double β-barrel fold, termed the “double jelly roll fold” (Figure 2) [18,19]. Interestingly, the double β-barrel fold is observed in other viruses that infect Bacteria (for example, PRD1 [20]) and Eukarya (for example, PBCV-1 [21]) (Figure 2). In addition to a homologous coat protein fold, these viruses also share other features, such as an internal lipid membrane and conserved genes (for example, a packaging ATPase) [14,22,23]. The cryo-EM structure of SH1, which infects hosts in Euryarchaeota, reveals that its coat protein adopts a single β-barrel fold [10]. It is therefore speculated that SH1 represents an ancient version of the coat protein fold that was then duplicated to generate the double β-barrel fold observed in STIV, PRD1, PBCV-1 and other viruses thought to be part of this lineage [10]. STIV and STIV2 are the only characterized viruses from this lineage that infect thermoacidophilic hosts. There are a variety of features of thermostability observed within the coat protein of STIV, such as its fold compactness and lack of cavities. However, other viruses from this lineage also contain such features and it is not known which of these specifically contribute to STIV's ability to maintain structural integrity within a thermophilic environment [19].

Certain viruses belonging to Lipothrixviridae and Rudiviridae have also demonstrated a conserved coat protein fold. In fact, a viral order (Ligamenvirales) encompassing both of these families has been proposed on the basis of genomic similarities between members of these two families that extend beyond conservation of the gene encoding the coat protein [9]. Viral members of both families are linear in morphology; however, those from Lipothrixviridae are enveloped and exist as (400–1,950) × 24–38 nm flexible, filamentous particles, while those from Rudiviridae are (610–900) × 23 nm non-enveloped, stiff, and rod-shaped in morphology [9]. Acidianus filamentous virus 1 (AFV1) belongs to Lipothrixviridae and encodes two structural proteins (orf132 and orf140), both of which contain an anti-parallel four-helix bundle fold that is structurally homologous to the C-terminal domain of the single MCP from Sulfolobus islandicus rod-shaped virus from Yellowstone National Park (SIRV-YNP) (Figure 3) [25]. Interestingly, tobacco mosaic virus, a thermostable, eukaryotic positive strand RNA virus, utilizes a heavily decorated 4-helix bundle to assemble a rod-like helical structure [26]. In contrast to tobacco mosaic virus, it is not yet known how these archaeal proteins are assembled into their presumably helical rod- or filamentous forms. However, a model has been proposed for coat protein assembly within AFV1: double-stranded DNA is thought to wrap around the positively charged protein AFV1 orf132, while AFV1 orf140 is proposed to interact with exterior of the DNA-protein bundle through its N-terminus and with the lipid membrane through its hydrophobic C-terminal helix [25].

Many of the published structures of archaeal viral proteins reveal DNA-binding motifs that are found in other organisms, such as the winged helix-turn-helix (wHTH) and the ribbon helix-helix (RHH) folds. The core of the wHTH fold is composed of a right handed three helix bundle followed by two β-strands, forming a 2 stranded antiparallel β-sheet known as the wing (Figure 4). The third helix of the three-helix bundle is the recognition helix, which inserts into the major groove of DNA [27], while the wing is often involved in nonspecific interactions with the ribose phosphate backbone. The wHTH motif is present in all three domains of life and often functions as a DNA recognition component of various transcription factors. The fold has also been found less frequently in proteins within RNA metabolism and those that are involved in protein-protein interactions [27]. Many archaeal virus proteins that adopt the wHTH motif are thus suggested to play roles in transcriptional regulation, although for many of these, specific binding sites within the viral or host genomes have yet to be characterized. Examples include F93 from STIV (unclassified viral family) and F93 from Sulfolobus spindle-shaped virus 1 (SSV1) (Fuselloviridae) (Figure 4). These dimeric wHTH proteins are structurally homologous to the prokaryotic MarR/SlyA protein family of transcription regulators [28,29] and are therefore expected to recognize (pseudo-) palindromic DNA targets.

The RHH fold is well named, beginning with a β-strand that precedes two α-helices. This β-strand interacts with the β-strand of another subunit, forming an anti-parallel two-stranded β-sheet that inserts into the major groove of DNA (Figure 5) [31]. While the wHTH motif is found in all domains of life, the RHH motif is found strictly in prokaryotes. Archaeal viral proteins that adopt the RHH fold (or elaborated versions of this) include E73 from Sulfolobus spindle-shaped virus from Ragged Hills (SSV-RH, Fuselloviridae) [32] and SvtR from SIRV1 (Rudiviridae) (Figure 5) [33]. SvtR is currently the best-characterized example of an archaeal virus protein containing the RHH motif. It is structurally homologous to bacterial RHH proteins and was determined to bind four target sequences from the SIRV1 genome. Target sequences include those preceding the coding region for its own gene as well as that for the coat protein; transcription of both the coat protein gene and its own gene were blocked by SvtR in an in vitro transcription assay [33].

To date, there have been five structurally characterized proteins that are thought to serve an enzymatic role in the context of a viral infection. Four out of five structurally characterized enzymes from archaeal viruses are involved in nucleic acid metabolism. These include crystal structures of two unrelated nucleases [34,35] as well as the SSV1 viral integrase (SSV1Int) [36,37] and a Rep protein (Figure 6) [38]. In addition, structural studies have also identified a putative gylcosyltransferase (Figure 7) [39]. Each of these enzymes is suggested to have a different evolutionary origin, highlighting the putative complexity and mosaicity of archaeal viral genomes.

Putative and characterized nucleases include SSV-RH D212 (Fuselloviridae) and AFV1 orf157 (Lipothrixviridae), respectively. SSV-RH D212 adopts the PD-D/EXK nuclease superfamily fold, however its activity remains uncharacterized and is therefore only a putative nuclease. This protein most closely resembles an archaeal Holliday junction resolvase. The full-length version is not active in the traditional Holliday junction cleavage assay and a clipped version only shows very low levels of metal-dependent nuclease activity, suggesting that the DNA binding surfaces are distinct between SSV-RH D212 and its closest homologs (Figure 6a) [35]. AFV1 orf157 very distantly resembles a trimmed down version of the two-layer sandwich fold of HIV1-integrase from the phosphonucleotidyl superfamily, which successfully guided experiments toward the functional characterization of orf157 as a nuclease (Figure 6b) [34].

The structurally and functionally characterized viral integrase from SSV1 (SSV1Int, Fuselloviridae) is a member of the tyrosine recombinase superfamily (Figure 6c). Interestingly, the biochemical [42] and structural analyses suggests that during strand exchange and cleavage, the enzyme assembles its active site in trans, which is consistent with mechanisms utilized by eukaryotic and not bacterial tyrosine recombinases [36,37]. SIRV1 orf119 (Rudiviridae) adopts a fold that is similar to superfamily II of Rep proteins, involved in the initiation of genomic replication (Figure 6c) [38]. Orf119 implements a novel “flip-flop” mechanism that is unique to conventional rolling circle replication (RCR) or rolling hairpin replication (RHR). This flip-flop mechanism may be employed by Rep proteins in other viruses that have closed, linear dsDNA genomes, such as members of the eukaryotic viral family, Poxviridae and bacteriophage N15 [38].

The crystal structure of STIV A197 reveals a putative glycosyltransferase adopting the GT-A fold whose closest structural homologs mainly consist of eukaryotic glycosyltransferases (Figure 7) [39]. Because the major coat protein of STIV is known to be glycosylated [14] and the virus assembles in the cytosol [43], A197 is a strong candidate for this activity, however it presently remains uncharacterized.